Elizabeth Woyke, writing for MIT Technology Review, tells us that UPS has come up with new ways to utilize machine learning and advanced analytics to handle the massive amounts of data required to move packages around the world and on time, regardless of the weather.
When snowstorms hit a major package processing center like Detroit, packages can be delayed as they travel through the bad weather towards their destinations. Now UPS is using an app called Network Planning Tools (NPT) to view all their facilities and divert their packages around the storms or move a particularly large shipment efficiently. Because of the large amounts of data, like weight, volume, and delivery deadlines that need to be analyzed in the decision-making process, UPS engineers are learning to rely on NPT to sort through how to solve their complex problems.
UPS is already using a system called ORION that helps by giving drivers their delivery routes and another program called EDGE that hones UPS’ internal processes.
NPT analyzes historical data to predict package volume and weight. The algorithms take into account past decisions made by UPS engineers against customer satisfaction and cost. From these data points, NPT can learn what options provide the best services.
UPS will deliver a whopping 800 million packages during the holidays this year. NPT can help pinpoint bottlenecks, like facilities with lagging package processing times and offer a solution within minutes that used to take weeks. UPS expects NPT to save them $100 to $200 million per year.
Reality Changing Observations:
Q1. In what other ways can UPS make packages move quickly and efficiently?
Q2. How can UPS show employees that they value their knowledge and while still prioritizing efficiency and building a collaboration between employees and their machines?
Q3. What are some other industries that could benefit from utilizing machine learning to handle large amounts of AI and predicting outcomes though algorithms?